🗺️ DMAP: A Distribution Map for Text

DMAP (ICLR 2026) is a mathematically grounded method that maps a text, via a language model, to a set of samples in the unit interval that jointly encode rank and probability information. This representation enables efficient, model-agnostic analysis and supports a range of applications.

DMAP works effectively with small evaluator language models such as OPT-125m that easily run on consumer hardware.

✨ Key Features

• 🎯 Intuitive Visualization: transform text into simple, informative, representations for downstream analysis
• 🔧 Easy Integration: Simple API that works with popular NLP libraries (transformers, scikit-learn, etc.)
• 📊 Rich Analytics: Built-in tools for quantitative and qualitative analysis of distribution patterns
• 🎨 Customizable: Easily plug-in new visualisations or analysis methods
• 📖 Interactive demo: Get up and running with DMAP in a few minutes

🚀 Quick Start

To install, simply run:

pip install git+https://github.com/Featurespace/dmap.git

Then, you may use DMAP as follows.

from dmap import DMAP

# Create and fit DMAP.
dmap = DMAP(evaluator_model='facebook/opt-125m')
text_map = dmap.fit(["The robot was dancing in the rain"])

# Visualize your DMAP samples.
dmap.plot()

For a more detailed example, we recommend cloning the repository and playing with our interactive demo.

📄 Citation

If you use DMAP in your research, please cite our paper accepted at ICLR 2026:

@article{dmap2025,
  title={DMAP: A Distribution Map for Text},
  author={Tom Kempton, Julia Rozanova, Parameswaran Kamalaruban, Maeve Madigan, Karolina Wresilo, Yoann Launay, David Sutton, and Stuart Burrell},
  year={2026},
  url={https://openreview.net/forum?id=SPElkPRurl}
}